I’ve been giving a number of presentations to executive and investors groups lately and one chart I compiled is perhaps one of the most interesting. Since I’m one of few how make an effort to track installed base of devices, I compiled a chart of all computing platform installed bases up to the current time period. Here is that chart.
There is a great deal of interesting observations I tease out in this chart, however, I want to focus on just one for the time being. Note the green part of the chart labeling the PC installed base. If we were to go back to the 2008-2009 timeframe and prior, we note that the vast majority of people using computers were on just one computing platform, the PC. Windows dominated this computing platform as the Mac’s annual market share average for the 2000-2010 time frame was 3%. While we knew we had many hundreds of millions of people using PCs the only real customer base segmentation we could do was to look at the split between corporate and consumer PCs. It wasn’t until 2008 when more consumer PCs shipped each year than corporate ones. Meaning within a base of many hundred millions we had a number of corporate PC users and consumer PC users and that was roughly all we know. That was the best we could segment the market. We knew some, but overall very little about user behavior in the PC user base at that time.
Fast forward to end of 2014 and the picture is quite different. The PC is not the only computing platform and now, thanks to smartphones, we have increased not only the number of computing devices but also the platforms as well. Why is this relevant. Look at the segmentation above the PC area of the graph. We have iOS, Google Android, and AOSP (non-Google Android) making up the vast majority of non-PC platform share. What is key about this observation is we know quite a bit about the behavior and even the economics of the nearly two billion people using one of those smartphone platforms. Fundamentally, what this more rich platform segmentation beyond the PC era has given us is a much deeper understanding of the behaviors of these platform customer bases than we ever had in the PC era. What has become glaringly clear is no customer base is the same in the three platforms which dominate computing usage today.
iPhone customer behave quite differently than Android customers. The economics, and the opportunity within both platforms are quite different. Similarly, since AOSP (non-Google Android) is primarily in China and we have learned and observed how different the China market is from the US, it highlights completely different opportunities than those with iOS and Google Android. But the key point is we are able to track the sizes of these customer bases. Which means we have a much easier time than ever before in attempting market sizing.
I joined Creative Strategies as an analyst to focus on consumer PC trends just before the dotcom bubble burst. I had been a part of several startups prior, and was briefed by many dotcom companies in my early days at Creative Strategies. Market sizing was one of the most difficult and often arbitrary tasks I observed startups attempt to do. Much of it had to do with such little knowledge of the behavior of the current PC user base at the time. Today this task has gotten a bit easier. If a company was to come to me today and articulate the type of customer they are going after, we can more cleanly estimate what the potential size of that market is by using our understanding of the different customers bases in each mobile platform and the behavioral differences within them.
Companies would like to believe their potential target market is all of the mobile installed base, however, it is simply just not true. Given what we are seeing with the differences of the behavior of iOS, Google Android, and non-Google Android, I am becoming increasingly convinced the opportunities, and strategies, within each segment are quite different. There is no one sized fits all solution for mobile. To address each platform requires almost as much nuance as addressing different global markets due to cultural, economic, political, and other factors which make each region so much different than the others. No global company can take the same strategy to the US, China, India, Latin America, etc., and from what I see the same need to alter a company strategy/business model for each global region, is also applicable to each computing platform.
Now that is a picture of the platform segmentation today but what about tomorrow? Here again this diagram of mine is helpful.
As we look to fill in the empty part of the diagram, will the Android customer base picture get clearer or more cloudy? If it is Google’s version of Android which dominates the market share in the below $200 phone segment then perhaps it may. However, if this diagram gets filled out with a variant of non-Google Android, perhaps a variant like Xiaomi, or Cyanogen, or something else, then we can continue to keep our platform segmentation into the next phase of mobile and seek to better understand the customer bases and the different opportunities within the different platforms taking share in the next billion users and beyond.
A lot to think about here! 🙂 [referring to your twitter post]
There is a big opportunity at the bottom of the pyramid. However, it may not look that way to the established players, who are making good money off of the top of the pyramid.
The bottom looks like a place where some company that is currently on the margins of mobile, and already dealing with that market, could do well. Just as one example, how about the m-pesa phone?
Yes m-pesa is a great example. Also FB may be very well positioned as more of an OS provider. Look at some of the deals they are doing in rural Africa, India, and others where it doesn’t cost to browse FB or some other sites. You could see these walled garden approaches working very well in this next phase. Even if it is built on Android, the challenge is Google’s services may be locked out due to the walled garden providers. It will be very interesting to watch this part play out.
But there is also still a great deal of innovation around mobile still to happen on the top of the pyramid. That is where understanding these customers bases becomes central.
What I find interesting about m-pesa is that, all you really need for it is SMS. You don’t need a smartphone. A feature phone is sufficient.
Although we tend to look at the opportunities that new technology opens up, there is often also a lot of opportunity with low-tech. For example, on the Japanese social network service Mixi, i-mode feature phone users overtook PC users in August, 2007, in terms of page-views (dark orange: i-mode feature phones, light-orange: PC). This illustrates how far you can go with feature phone technology.
http://www.garbagenews.com/img/gn-20081110-09.jpg
Looking at Google and Facebook, you can see their different approaches to technology. Google’s initiatives like project Loon and Android One aim to improve the underlying Internet technology platform. On the other hand, Facebook’s projects like Internet.org aim to make the most of the platforms that are already available. Facebook’s approach is, at least short term, vastly more practical.
I often think back to the beginning of the commercial Internet. Instead of waiting for the Information Superhighway, Netscape bet that analog telephone lines would be enough. I think that this kind of scrappy approach is often what you need.
Although I agree that Facebook has an inherent advantage because it can take a walled garden approach, I think Google could have worked harder to squeeze the last bit of performance of what was already there. For example, Google could have done what Opera did with Opera Mini, instead of pouring resources into a browser that is often accused of being very resource intensive. Opera Mini often compresses HTML pages by more than 90%, but it sacrificing compatibility with cutting-edge Javascript.
Therein however, lies the conflict within Google and Android. Targeting the developed countries with the Chrome browser, while simultaneously targeting emerging countries with a Opera Mini-like browser creates an internal conflict.
“What I find interesting about m-pesa is that, all you really need for it
is SMS. You don’t need a smartphone. A feature phone is sufficient.” Yes. I would think, though, that m-pesa is in a very good position to sell inexpensive smartphones to their users. They could offer financing of the smartphone purchase over the phone.
I totally agree. In the emerging countries, my understanding is that the player who are making money are 1) payments, 2) carriers, and 3) e-commerce. It stands to reason that either of them will try to control the smartphone ecosystem in that country, and offering their brand of smartphone would be one way to do that. What is important is the absence of advertising as a money maker, and hence the limited control that Google can exert.
The question then becomes, what will the “inexpensive smartphone” look like? Will it be a Google Android or will it be AOSP Android? It could even be a Tizen, Firefox OS or Windows Phone. I would argue that it could even be a Symbian, maybe running Opera Mini in a Mobile Java environment.
Yes, Facebook in Africa, that’s fascinating. Do they have the will and the knowledge of the local markets to make customers happy? Can they make money doing so? It will be interesting to find out.
Innovation at the top – do you mean people will be doing different things with their phones in the future, or doing the same things in different ways? The different things I can think of happen because of sensors worn on the body, with the phone acting as “digital hub.”
Looking at the top of the pyramid, Apple’s revisions to their phones last year were all about strengthening their ownership of the top 1/3 tier. And it’s looking like the competition is largely ceding that tier to Apple. Which suggests to me that in the not too distant future, as they start to run out of non-customers who can afford to become their customers, they’re going to be looking harder at ways they can own more of the middle tier.
Given how the 5c did, they may not want to make a “cheaper” phone — but what they could perhaps do, when they decide to gain more customers at the top end of the mid-range phone market, is change their discounting regime for older phones. Instead of the 2 year old model being $200 cheaper than the current model, they could make it $300 cheaper (with the 1 year old phone being $150 below the current model). That would probably drive the midrange android even further downmarket, and they’d be able to own the top 1/2 of the pyramid instead of the top 1/3.
Another possibility would be to keep the 3-year old phone around (as they are doing now in emerging markets like China), in more affluent markets as well. Easy enough to distinguish the two price grades of “free on contract” models by making the older one be “free with the first month free” or some such.
Samsung does seem to be doubling down on high price, high specs and design though. We’ll see if that pays off.
It’s kind of funny to see Premium starting at $400 nowadays, used to be $600 or $500 not too long ago.
There was also a time when pretty much any smartphone was premium.
Joe
Actually, I’ve been wondering about that all day, but from the other end up: are smartphones the new dumbphones ?
– with current prices, buying a dumbphone doesn’t really make sense, to save $50 max and get a piddly screen.
– these late adopters are not really tech-motivated, mostly very old or very young
– increasingly, I see the smart part of smartphones not being used much if at all: no media, barely any apps. Symbian could play music, videos, WAP and bathroom games (Candy Crush would run on S40), take photos… I’d guess 30-50% of users are not doing more than that.
– I’m even seeing the young go back to SMS and MMS now that these are for free on contract & the UI kind of emulates group chats.
I’d be really interested on info on that. And on which apps do break through to unmotivated users.
Ben,
Could you humor me and give a little insight into your definition of “installed base?”
Knowing that computer sales are not cumulative (so many sales are replacements for devices that will not be re-used), the devil-in-details of your methodology would be enlightening.
Thanks.
It is devices in use in the market. I track devices being actively used in the market currently.
Um. Yeah. I was hoping for more info on how it is that you “track” them, what constitutes “active,” and what assumptions you employ in your estimates.
Thanks.